Upper motor neurons are a target for gene therapy and UCHL1 is necessary and sufficient to improve cellular integrity of diseased upper motor neurons.

There are no effective cures for upper motor neuron (UMN) diseases, such as amyotrophic lateral sclerosis (ALS), primary lateral sclerosis, and hereditary spastic paraplegia. Here, we show UMN loss occurs independent of spinal motor neuron degeneration and that UMNs are indeed effective cellular targets for gene therapy, which offers a potential solution especially for UMN disease patients. UCHL1 (ubiquitin C-terminal hydrolase-L1) is a deubiquitinating enzyme crucial for maintaining free ubiquitin levels. Corticospinal motor neurons (CSMN, a.k.a UMNs in mice) show early, selective, and profound degeneration in Uchl1nm3419 (UCHL1-/-) mice, which lack all UCHL1 function. When UCHL1 activity is ablated only from spinal motor neurons, CSMN remained intact. However, restoring UCHL1 specifically in CSMN of UCHL1-/- mice via directed gene delivery was sufficient to improve CSMN integrity to the healthy control levels. In addition, when UCHL1 gene was delivered selectively to CSMN that are diseased due to misfolded SOD1 toxicity and TDP-43 pathology via AAV-mediated retrograde transduction, the disease causing misfolded SOD1 and mutant human TDP-43 were reduced in hSOD1G93A and prpTDP-43A315T models, respectively. Diseased CSMN retained their neuronal integrity and cytoarchitectural stability in two different mouse models that represent two distinct causes of neurodegeneration in ALS.

INTRODUCTION

Motor neuron circuitry is one of the most complex circuitries in our body; it has important cellular and neuronal components both in the motor cortex and in the spinal cord and is responsible for the initiation and modulation of voluntary movement [1-3]. In motor neuron diseases, such as hereditary spastic paraplegia (HSP), primary lateral sclerosis (PLS) and amyotrophic lateral sclerosis (ALS), this complex circuitry degenerates [4-13]. There has been a long debate in the field about relevant contribution of different components of the circuitry so that targeted therapies can be developed [14-21]. Many argued that corticospinal motor neuron (CSMN) or the upper motor neuron (UMN) degeneration is a byproduct or a consequence of the “die-back” phenomenon [14-16]. This hypothesis unfortunately eliminated their importance and diminished enthusiasm of targeting UMNs as a cellular target even for UMN diseases [15].

Recent building evidence, however, began to reveal that UMNs degenerate early in diseases and the cortex is indeed a proper target for therapeutic interventions [1, 17–27]. Suppression of the mutant SOD1 only in the motor cortex was enough to delay disease onset in SOD1G93A rat model of ALS and extended their survival [28]. Transplantation of neural progenitor cells expressing GDNF into the motor cortex in SOD1G93A rat model of ALS provided neuroprotection in both motor cortex and spinal cord, delayed disease pathology and extended life span [29]. Cre-lox mediated ablation studies also confirmed that CSMN degeneration occurs via cell-autonomous mechanisms [30, 31]. In addition, cortical hyperexcitability is used as an early detection marker in ALS patients [1, 32–34]. Therefore, building evidence now reveal the importance of CSMN and UMN to disease pathology, and that they indeed are cellular targets for therapeutics [23, 24, 35].

UCHL1 is an important component of the ubiquitin-proteasome system (UPS) and can either add to or remove ubiquitin from polyubiquitin chains.[36-38] Mutations in UCHL1 gene cause autosomal recessive spastic paraplegia-79 (SPG79) [39-43]. Patients develop spasticity with involvement of UMNs [39]. The UCHL1−/− mice, which lack all UCHL1 function [44], display motor function defects and progressive CSMN loss that accompanied by vacuolated apical dendrites, spine loss, and increased ER stress, which starts very early [45]. The spinal motor neurons (SMN) in these mice are also affected, and display disintegration of neuromuscular junctions (NMJ) [46]. Therefore, the same question arises: do CSMN degenerate because SMN are unhealthy? In an effort to develop gene therapy approaches selectively to UMNs, one must reveal whether CSMN degeneration is a function of SMN health, or whether CSMN health can be improved independently of SMN. Rbp4creUCHL1f/f and HB9creUCHL1f/f mice, which lack UCHL1 function only in large subcerebral projection neurons (SCPN) in layer 5, –which includes CSMN–, or in SMN in the spinal cord, offer great advantage to address this important issue. We find that CSMN degeneration is not a function of SMN health and UMNs are indeed valid and effective cellular targets for gene therapy.

AAV2-mediated UCHL1 gene delivery only to the CSMN of UCHL1−/− mice was sufficient to improve the cytoarchitectural integrity and stability of diseased CSMN, such that they become comparable to healthy CSMN. These results further prove the importance of UCHL1 for the health and integrity of UMNs and suggests UCHL1 as a potential candidate for gene therapy for diseased UMNs. One of the cellular pathologies that is shared among many UMNs that become diseased due to different underlying causes, such as misfolded SOD1 toxicity [47-50], TDP-43 pathology [51, 52], lack of Alsin function [53], or mutations in the profilin gene [54], is disintegration of apical dendrite and spine loss. Interestingly, exactly the same pathology is detected in the Betz cells of a broad spectrum of ALS patients, including sALS, fALS and ALS/FTLD [9], UMNs in two different species display same cellular defects, suggesting that translation is at a cellular level and that improvement of CSMN health and stability will translate to UMN improvement in patients.

Here, we find that AAV-mediated UCHL1 expression selectively in CSMN that are diseased due to mSOD1 toxicity and TDP-43 pathology, recovers their disease state, improves their neuronal integrity and stability, reduces their protein aggregation such that they become comparable to CSMN of healthy controls. hSOD1G93A and prpTDP-43A315T mice are one of the best characterized mouse models of ALS [47-52], and they recapitulate many aspects of ALS disease pathology, including progressive CSMN loss [47, 48, 51]. Yet, they represent different underlying causes of ALS and motor neuron degeneration [55-59].

Our findings show further proof that UMNs are indeed proper cellular targets for gene therapy approaches, especially for UMN diseases, and that their degeneration is not a function of SMN health. We also show evidence that direct delivery of UCHL1 to CSMN that become diseased due to mSOD1 toxicity and TDP-43 pathology is sufficient to improve their cytoarchitectural integrity and apical dendrite stability. Our results reveal UCHL1 as a potential target for gene therapy approaches for diseased UMNs in ALS and other related motor neuron diseases.

MATERIALS and METHODS

Mice

All animal procedures were approved by Northwestern University Animal Care and Use Committee and conformed to the standards of the National Institutes of Health. Uchl1 (UCHL1−/−) mice carry a spontaneous 795 base-pair intragenic deletion that results in the removal of 24 base-pairs of exon 6 and 771 base-pairs of intron 6 [44-46]. Heterozygous mice (UCHL1+/−) were viable, fertile, and bred to generate UCHL1 deficient (UCHL1−/−) mice. Survival times and motor function defects were comparable between males and females with 100% penetrance. Uchl1tm1a(EUCOMM)Hmgu knockout first allele targeted embryonic stem cells were purchased from The European Conditional Mouse Mutagenesis (EUCOMM) Program. Floxed UCHL1 mice, in which exon 4 of the UCHL1 gene is flanked by LoxP sites, were generated with the assistance of Northwestern University Transgenic and Targeted Mutagenesis Laboratory. The lacZ/neo cassette flanked by FRT sites in intron 3 was deleted by crossing germline transgenic mice with EIIa FLPeR mice (provided by Northwestern University Transgenic and Targeted Mutagenesis Laboratory), to generate the UCHL1f/f mice. Rbp4cre mice on a mixed background [Tg(Rbp4-cre)KL100Gsat/Mmucd; MMRRC stock# 031125-UCD] were purchased from Mutant Mouse Regional Resource Center (MMRRC) at UC Davis Mouse Biology Program [60-65]. HB9cre mice [B6.129S1-Mnx1/J; JAX stock #006600] were purchased from Jackson labs [66, 67]. Both floxed UCHL1 and Rbp4cre mouse lines were backcrossed to C57BL/6J background for at least 8 generations. Conditional mutant mice were generated by crossing floxed UCHL1 mice with Rbp4cre or HB9cre mice. hSOD1G93A mice expressing the high copy number of human SOD1 with the G93A point mutation were purchased from Jackson Labs (B6SJL-Tg(SOD1*G93A)1Gur/J; JAX stock # 002726) and backcrossed to C57BL/6J background in our laboratory [48, 50]. prpTDP-43A315T mice expressing the human TDP-43 with the A315T point mutation under the control of the prp promoter were purchased from Jackson Labs (B6.Cg-Tg(Prnp-TARDBP*A315T)95Balo/J; JAX stock # 010700) [51, 52]. Primers used to determine genotype of Uchl1 mice are UCHL1 forward: tggacggctgtgtgtgctaatg, WT reverse: ctaagggaagggtcttgctcatc, mutant (Mt) reverse: gtcatctacctgaagagagccaag, yielding 668bp WT and 334bp Mt PCR products. Primers used to determine genotype of floxed UCHL1 mice are forward: tagtccaatccttgtaccagttgg and reverse: ccatggttctagatgctgttgaatgc, yielding 428 bp WT and 540 bp floxed UCHL1 products. Primers used to determine genotype of cre mice are forward: gcattaccggtcgatgcaacgagtgat and reverse: gagtgaacgaacctggtcgaaatcagt, yielding a 408 bp product. Primers used to determine genotype of hSOD1G93A mice are forward: catcagccctaatccatctga and reverse: cgcgactaacaatcaaag, yielding a 236 bp product. Primers used to determine genotype of prpTDP-43A315T mice are forward: ggatgagctgcgggagttct and reverse: tgcccatcataccccaactg, yielding a 400 bp product.

Generation of adeno-associated virus (AAV)

AAV vectors were generated by the University of Pennsylvania Vector Core facility by triple transfection of subconfluent HEK293 cells using three plasmids: an AAV trans-plasmid encoding AAV2 capsid, an adenovirus helper plasmid pΔF6, and an AAV cis shuttle plasmid expressing eGFP driven by a CMV promoter (pENN.AAV.CMV.PI.eGFP.WPRE.bGH). The culture medium was collected, concentrated by tangential flow filtration and purified by iodixanol gradient ultracentrifugation as previously described [68]. pGEM-T vector plasmid containing the mouse UCHL1 cDNA ORF clone was purchased from Sino Biological (cat: MG50690-G, Wayne, PA, USA), and the UCHL1 CDS was subcloned into a AAV plasmid with CBA promoter [69], to generate pAAV.CBA.UCHL1-IRES-eGFP.WPRE plasmid that was packaged into AAV2 virus particles by the University of Pennsylvania Vector Core facility as described above.

Retrograde labeling and transduction surgeries

Surgeries were performed on a stereotaxic platform. Micro-injections were performed using pulled-beveled glass micro-pipettes attached to a nanojector (Drummond Scientific, Broomall, PA, USA). CSMN were retrogradely labeled by AAV encoding eGFP (AAV2-eGFP: 96 nl containing 4.88 × 108 genome copies; AAV2-UCHL1-IRES-eGFP: 414 nl containing 3.39 × 109 genome copies), injected into the CST and CSMN were retrogradely transduced as described [45, 70]. Briefly, a small laminectomy at the cervical spinal cord (C2-C3) level was performed to expose the spinal cord, and to study cytoarchitecture of individual CSMN, a low dose of AAV is injected into the CST that lies within the dorsal funiculus (df) at 0.3 mm depth. This allows visualization of a subset of CSMN without too much overlap with neighboring cells, in a pattern similar to Golgi’s silver stain, enabling detailed analysis of their morphology. To visualize maximum number of CSMN, 0.5% Fluoro-Gold (FG; Fluorochrome LLC, Denver, CO; 96nl in 0.9% saline solution) was injected as described. Mice were randomly assigned to receive either the AAV2-eGFP or AAV2-UCHL1-IRES-eGFP injections.

Rotarod, hanging wire, and grip strength tests

Rotarod: A rotating rod that accelerates linearly from 4 to 40 rpm (Ugo Basile, Gemonio, VA, Italy) was used and an average time spent on the rotating rod for three consecutive trials was calculated for maximum 5 min. Hanging wire: Time each mouse spent hanging on to an upside-down wire mesh (50 cm above a bench) was recorded in 3 consecutive trials. Grip strength: Peak force applied by mice (forelimbs and hindlimbs recorded separately) was measured using grip strength meter (Ugo Basile, Gemonio, VA, Italy) in 3 consecutive trials.

Tissue collection and histology

Mice were deeply anesthetized with ketamine (90 mg/kg) and xylazine (10 mg/kg) and perfused with 4% PFA in PBS. The brain was removed intact from each mouse, post-fixed by 4% PFA overnight and kept in PBS-sodium azide (0.01%) at 4°C. Brains were sectioned (coronal; 50 μm) using a vibrating microtome (VT1000S, Leica Instruments, Nussloch, Germany)

Immunocytochemistry

Immunocytochemistry was performed on every 6th coronal section of mouse brains. Antigen retrieval was performed for Ctip2 immunocytochemistry; sections were treated with 0.01 M sodium citrate, pH 9.0, at 80°C water bath for 2 hr prior to incubation with primary antibody. Primary antibodies were: anti-Ctip2 (1:500; Abcam ab18465, Cambridge, MA, USA); anti-GFP (1:1000; Invitrogen A11122, Grand Island, NY, USA), anti-ChAT (1:200; Millipore AB144P, Burlington, MA, USA), anti-misfolded SOD1 (B8H10, 1:250, Médimabs MM-0070-P, Montreal, Quebec, Canada), anti-FLAG clone M2 (1:500, Sigma F1804, St. Louis, MO, USA), and anti-UCHL1 (1:1000; ProteinTech 14730-1-AP, Rosemont, IL, USA). After PBS washes, either fluorescent conjugated (AlexaFluor, Invitrogen, Grand Island, NY, USA) or biotinylated (Vector Laboratories, Burlingame, CA), USA secondary antibodies were used for detection.

Quantification of misfolded SOD1 and TDP-43

50 μm thick free-floating sections of primary motor cortex were immunostained with the same antibody dilution at the same time. B8H10 monoclonal antibody can detect a wide spectrum of SOD1 mutants and metal-depleted WT SOD1 protein, but not intact WT SOD1 [71]. prpTDP-43A315T mice express the human TDPA315T protein with an N-terminal FLAG tag, therefore anti-FLAG antibody can be used to selectively visualize the exogenous TDP-43A315T transgene product [52]. All images were captured using the 20X objective on an inverted epifluorescent Eclipse TE2000-E microscope (Nikon Inc., Melville, NY, USA) using the same settings and exposure time for all samples. Retrogradely labeled eGFP positive (eGFP+) UMNs were traced using Image J (NIH, Bethesda, MD, USA). ROI were transferred to the B8H10 or FLAG channel, and the mean gray value was measured to determine the levels of misfolded SOD1 or human TDPA315T protein in GFP+ UMNs only. All retrogradely labeled UMNs were included in the analysis, only UMNs with overlapping B8H10 or FLAG signal coming from other cells in other focal planes were excluded. Average mean gray value ± SEM was reported per experimental group.

Imaging, quantification, and statistical analysis

Nikon SMZ1500 and Nikon Eclipse TE2000-E fluorescence microscopes equipped with Intensilight C-HGFI (Nikon Inc., Melville, NY, USA) were used. Epifluorescence images were acquired using a Digital Sight DS-Qi1MC CCD camera (Nikon Inc., Melville, NY, USA) and light images were acquired using a Ds-Fi1 camera (Nikon Inc., Melville, NY, USA). Confocal images were collected using a Zeiss 510 Meta confocal microscope (Carl Zeiss Inc., Thornwood, NY, USA). Numbers and diameters of CSMN were determined at P100. Since the AAV is injected into the CST and motor cortex is analyzed, the only cells that are labeled with this method are CSMN in layer 5 of primary motor cortex. Therefore, all cells with intact cell bodies were included in analysis regardless of their specific coordinates. Average CSMN diameter (at least 100 neurons / mouse; n = 3) was measured using Elements Software (Nikon Inc., Melville, NY, USA) OR Image J Software (NIH, Bethesda, MD, USA). For CSMN spine density measurements, a measured segment of apical dendrites (in layer 2/3, in the primary apical dendrite) were selected and the total numbers of spines were counted in each segment (10 segments/mouse; n = 3). The numbers were averaged, and results were presented as average number of spines per μm, per genotype. Even though all statistical analyses were performed per genotype, the information of the total counts are included in the results section to be informative on the extent of the quantitative assessment. All quantifications were performed blinded to the genotype and treatment of the mice.

Statistical analyses were based on the average numbers for each mouse, and not based on total individual number of counts. All statistical analyses were performed using Prism software (version 5.0a; Graphpad Software Inc., La Jolla, CA, USA). Statistically significant differences were determined after either one-way ANOVA with post hoc Tukey’s multiple comparison tests or t-test. Statistically significant differences were considered at p < 0.05, and values were expressed as the mean ± SEM.

RESULTS

Site-specific in vivo deletion of UCHL1

To investigate the impact of the spinal motor neuron (SMN) dysfunction on CSMN cellular defects, and to assess the cell-autonomous importance of UCHL1 function for CSMN, we generated two conditional mutant mice, which lacked UCHL1 function either in CSMN or in SMN. First the floxed UCHL1 mice (UCHL1f/f), in which the exon 4 of the Uchl1 gene is flanked by loxP sites, were generated (Fig. 1a). In the presence of cre recombinase, two loxP sites in introns 3 and 4 are recombined leading to deletion of Uchl1 exon 4. This deletion introduces a de novo stop codon shortly after exon 3 in the Uchl1 open reading frame (Fig. 1b), eliminating the production of a functional UCHL1 protein. Therefore, when UCHL1f/f mice were crossed with Rbp4cre mice, in which cre recombinase is expressed under the control of the Rbp4 promoter targeting SCPN (subcerebral projection neurons) in layer 5, –including the CSMN in the primary motor cortex–[60-63], UCHL1 is deleted from the CSMN in Rbp4cre UCHL1f/f mice (Fig. 1c). UCHL1f/f mice were also crossed with HB9cre [66, 67, 72–75] to generate the HB9cre UCHL1f/f mice, which lack UCHL1 protein in SMN but not in CSMN (Fig. 1c). These transgenic mice are generated to determine whether depletion of UCHL1 function in the SMN would have an impact on the health of CSMN.

Fig. 1:

UCHL1 expression can be deleted using cre/lox conditional mutant approach in the CSMN of floxed UCHL1 mice.

a Floxed UCHL1 (UCHL1f/f) mice were generated by introducing loxP sites flanking the exon 4 of UCHL1. Cre recombinase activity leads to removal of the exon 4. b Deletion of exon 4 of mouse UCHL1 gene introduces a de novo stop codon shortly after the deletion site. c Rbp4cre mice drives the expression of cre recombinase in layer 5 subcortical projection neurons including the CSMN. HB9cre mice drives the expression of cre recombinase in spinal motor neurons located in the ventral horn of the spinal cord. Crossing these cre driver mice with UCHL1f/f mice leads to conditional mutant mice that lack UCHL1 function either in the CSMN or the SMN. d-h Representative images of primary motor cortex from UCHL1f/f (d-e), Rbp4cre UCHL1f/f (f-g), and HB9cre UCHL1f/f (h) mice at P30 (d, f) and P100 (e, g, h). CSMN can be identified by Ctip2+ nuclei. Scale bars on top panel = 250 μm d’-h’ Representative images of layer 5 of primary motor cortex from UCHL1f/f (d’-e’), Rbp4cre UCHL1f/f (f’-g’), and HB9cre UCHL1f/f (h’) mice at P30 (d’, f’) and P100 (e’, g’, h’). Scale bar = 50 μm. d”-h” Representative images of CSMN from UCHL1f/f (d”-e”), Rbp4cre UCHL1f/f (f”-g”), and HB9cre UCHL1f/f (h”) mice at P30 (d”, f”) and P100 (e”, g”, h”) captured by confocal microscope. Scale bar = 10 μm.

CSMN can be identified their position in the layer 5 of primary motor cortex and expression of Ctip2 [2, 45, 48, 76, 77]. As expected, Ctip2+ CSMN that normally have high levels of UCHL1 protein were not affected in the UCHL1f/f mice (Fig. 1d–d’, e–e’), as CSMN continued to express endogenous UCHL1 at both postnatal day P30 and P100. However, even though UCHL1 expression was detected in other neurons within the motor column, CSMN completely lacked UCHL1 protein in the Rbp4cre UCHL1f/f mice (Fig. 1f–f’, g–g’). In contrast, CSMN of HB9cre UCHL1f/f mice continued to express endogenous levels of UCHL1 and their levels were comparable to that of UCHL1f/f mice (Fig. 1h–h”).

UCHL1 levels were not affected in SMN of UCHL1f/f mice (Fig. 2a–a’, b–b’). In contrast, SMN of HB9cre UCHL1f/f mice lacked UCHL1 expression at both P30 and P100 (Fig. 2c–c’, d–d’). Especially the large alpha motor neurons were devoid of UCHL1 expression in the ventral horn of the spinal cord (Fig. 2c’–d’). However, the SMN of the Rbp4cre UCHL1f/f mice continued to express UCHL1 even at P100 (Fig. 2e and e’), further confirming selective and robust deletion of UCHL1 in CSMN of the Rbp4cre UCHL1f/f mice and in SMN of HB9cre UCHL1f/f mice. It is also important to note that the absence of UCHL1 function did not affect the birth and specification of neither CSMN nor the SMN, as Ctip2+ CSMN and ChAT+ SMN were detected in both of these conditional mutant lines.

Fig. 2:

HB9cre UCHL1f/f mice lack UCHL1 in their spinal motor neurons (SMN)

a-e Representative images of the ventral horn of the spinal cord of UCHL1f/f (a-b), HB9cre UCHL1f/f (c-d), and Rbp4cre UCHL1f/f (e) mice at P30 (a, c) and P100 (b, d, e). Scale bar = 100 μm. a’-e’ Representative images of ChAT+ SMN from UCHL1f/f (a’-b’), HB9cre UCHL1f/f (c’-d’), and Rbp4cre UCHL1f/f (e’) mice at P30 (a’, c’) and P100 (b’, d’, e’) captured by confocal microscope. Scale bar = 20 μm.

CSMN defects in the absence of UCHL1 function are cell autonomous

CSMN were retrogradely transduced by AAV-2 eGFP, via injection into the corticospinal tract at P30 [2, 45, 70] and were analyzed at P100 (Fig. 3a). CSMN of UCHL1f/f control mice had large pyramidal cell bodies, prominent apical dendrites, and spines throughout the dendrites (Fig. 3b–d). The CSMN soma size (Fig. 3f) in UCHL1−/− (12.59 ± 0.22 μm, n = 219 CSMN; n = 9 mice) and Rbp4cre UCHL1f/f mice (13.35 ± 0.17 μm, n = 469 CSMN; n = 9 mice) were both significantly reduced when compared to the CSMN of UCHL1f/f mice (14.44 ± 0.19 μm, n = 457 CSMN; n = 9 mice; UCHL1f/f vs. UCHL1−/− p = 0.0007; UCHL1f/f vs. Rbp4cre UCHL1f/f p = 0.0052), but they were comparable to each other (p = 0.2941). Interestingly, CSMN soma size in HB9cre UCHL1f/f mice (14.37 ± 0.21 μm, n = 440 CSMN; n = 13 mice) was comparable to that of CSMN in UCHL1f/f (p = 0.9944), suggesting that deletion of UCHL1 in the SMN did not have impact on the overall health of CSMN.

Fig. 3:

Cortex specific UCHL1 conditional mutant mice recapitulate pathology observed in the CSMN of UCHL1−/− mice.

a Schematic overview of the experimental setup. CSMN of UCHL1f/f, UCHL1−/−, Rbp4cre UCHL1f/f, and HB9cre UCHL1f/f mice were retrogradely transduced by injecting AAV2-eGFP into the corticospinal tract (CST) at P30, and CSMN were analyzed at P100. b Representative images of GFP labeled CSMN after DAB mediated immunohistochemistry enhancement of the GFP signal allowing investigation of CSMN morphology in detail. Scale bar = 20 μm. c Primary apical dendrites of CSMN retrogradely labeled by AAV2-eGFP. Scale bar = 10 μm. d Cell bodies of CSMN retrogradely labeled by AAV2-eGFP. Scale bar = 20 μm. e Quantitative analysis of apical dendrites reveals significant increase in the average percentage of apical dendrites with vacuoles in the absence of UCHL1. Data presented as mean ± SEM; *p < 0.05; **p < 0.01; ***p < 0.001; one-way ANOVA with post hoc Tukey’s multiple comparison test. f Quantitative analysis of soma size shows significant reduction in the average CSMN diameter in the absence of UCHL1. Data presented as mean ± SEM; **p < 0.01; ***p < 0.001; one-way ANOVA with post hoc Tukey’s multiple comparison test.

We next investigated whether the integrity of apical dendrites was equally compromised in Rbp4cre UCHL1f/f and HB9cre UCHL1f/f mice. In line with previous reports, UCHL1−/− mice (81.16 ± 3.31%, n = 357 apical dendrites, n = 3 mice) exhibited significantly higher percentage of vacuolated CSMN apical dendrites (Fig. 3e) than CSMN of UCHL1f/f mice (26.08 ± 7.16%, n = 654 apical dendrites, n = 9 mice, p = 0.0011). Majority of CSMN in Rbp4cre UCHL1f/f mice also had vacuolated apical dendrites (52.82 ± 7.29%, n = 701 apical dendrites, n = 11 mice; p = 0.0235). Interestingly, CSMN of UCHL1−/− and Rbp4cre UCHL1f/f mice were comparable (p = 0.1401). In contrast, CSMN of HB9cre UCHL1f/f mice, –which lacked UCHL1 only in the spinal cord, leaving UCHL1 intact in the brain–, displayed healthier morphology and percent apical dendrites with vacuoles were comparable to the CSMN of UCHL1f/f mice (23.68 ± 8.68%, n = 842 apical dendrites, n = 13 mice; p = 0.992).

Spine numbers were reduced in apical dendrites that disintegrate. Average number of spines per μm of healthy primary apical dendrite (Fig. 4a) of the CSMN in UCHL1f/f (0.70 ± 0.15 spines/μm, n = 3 mice), UCHL1−/− (0.63 ± 0.14 spines/μm, n = 3 mice), Rbp4cre UCHL1f/f (0.87 ± 0.02 spines/μm, n = 3 mice), and HB9cre UCHL1f/f mice (0.63 ± 0.07 spines/μm, n = 3 mice) were comparable at P100 (Fig. 4b). Similarly, the average number of spines per μm of the vacuolated and disintegrating primary apical dendrite of the CSMN in UCHL1f/f (0.37 ± 0.01 spines/μm, n = 3 mice), UCHL1−/− (0.35 ± 0.07 spines/μm, n = 3 mice), Rbp4cre UCHL1f/f (0.38 ± 0.15 spines/μm, n = 3 mice) and HB9cre UCHL1f/f mice (0.33 ± 0.03 spines/μm, n = 3 mice) were similar (Fig. 4c). However, diseased apical dendrites always had reduced number of spines than healthy apical dendrites regardless of genotype. Since there is strong correlation between spine loss and the integrity of the apical dendrite [1, 9, 45, 51, 70, 78–83], for future quantitative assessments, we focused our attention to the health and integrity of the apical dendrite for each genotype and experimental condition.

Fig. 4:

Spine density is reduced in degenerating and vacuolated apical dendrites of CSMN.

a Representative images of healthy and vacuolated CSMN primary apical dendrites at P100. Scale bar = 10 μm. b Quantitative analysis of average spine density along the primary apical dendrites of healthy CSMN. Data presented as mean ± SEM, one-way ANOVA with post hoc Tukey’s multiple comparison test. c Quantitative analysis of average spine density along primary apical dendrites of CSMN with vacuoles. Data presented as mean ± SEM, one-way ANOVA with post hoc Tukey’s multiple comparison test.

UCHL1 delivery to CSMN of UCHL1−/− mice is sufficient to improve CSMN integrity

We next took advantage of AAV2 mediated gene delivery to restore UCHL1 function only in the CSMN of UCHL1−/− mice (Fig. 5). AAV2 containing an UCHL1-IRES-eGFP bicistronic expression vector was injected into the corticospinal tract (CST) of UCHL1−/− mice at P30 and CSMN were analyzed at P100 (Fig. 5a). Wild type (WT) and UCHL1−/− mice that underwent same retrograde transduction surgery using the AAV2-GFP vector alone that does not contain the UCHL1 coding sequence were used as positive and negative controls, respectively. WT CSMN transduced by AAV2-eGFP express endogenous UCHL1 (Fig. 5b,e). UCHL1 protein is not detected in UCHL1−/− mice, and CSMN that are transduced with AAV2-eGFP also lack UCHL1 expression (Fig. 5c,f). In contrast, when CSMN are retrogradely transduced with AAV2-UCHL1-IRES-eGFP, which leads to the expression of both GFP and UCHL1 proteins, high level of UCHL1 expression is detected only in GFP+ transduced CSMN (Fig. 5d,g), confirming effective transduction and directed gene delivery only to CSMN. AAV2-UCHL1-IRES-eGFP induces both UCHL1 and eGFP expression (not a fusion protein due to the bicistronic expression vector containing the IRES between two coding sequences), and because eGFP expression can be detected throughout the neuron, the integrity of the apical dendrites can be precisely assessed (Fig. 5b–d). Apical dendrites of WT mice were intact (Fig. 5b), but the apical dendrites of UCHL1−/− mice were mostly filled with vacuoles (Fig. 5c). Interestingly, upon direct UCHL1 delivery to the CSMN, the integrity of the apical dendrites was dramatically improved; vacuolization of the apical dendrites was reduced and at times completely eliminated (Fig. 5d).

Fig. 5:

UCHL1 expression can be restored in CSMN of UCHL1−/− mice using a AAV2-UCHL1-IRES-eGFP bicistronic expression vector.

a Schematic overview of the experimental setup. CSMN of wild type (WT), or UCHL1−/− mice were retrogradely transduced by injecting AAV2-eGFP into the corticospinal tract (CST) at P30, and CSMN were analyzed at P100. AAV2-UCHL1-IRES-eGFP virus was used for retrograde transduction of CSMN with UCHL1 expression in the UCHL1−/− mice. b-g Representative images of CSMN apical dendrites (b-d), and soma (e-g) retrogradely transduced by AAV2-eGFP in WT (b,e), UCHL1−/− (c,f) and AAV2-UCHL1-IRES-eGFP in UCHL1−/− (d,g) mice. Scale bar b-d = 20 μm, e-g = 10 μm.

We used anti-GFP immunohistochemistry with DAB substrate to amplify GFP signal allowing precise analysis of retrogradely transduced CSMN (Fig. 6, Supplementary Fig. 1a). Using a low dose of AAV for retrograde transduction, enabled labeling only a small subset of CSMN with minimal overlap, –generating a pattern similar to Golgi’s silver stain–, allowing detailed visualization and investigation of their cellular morphology with precision [2, 45, 70]. CSMN of WT mice displayed healthy robust apical dendrites (22.57 ± 3.48 %, n = 735 apical dendrites, n = 5 mice), whereas CSMN of UCHL1−/− mice contained significantly higher percentages of vacuoles (81.16 ± 3.31 %, n = 357 apical dendrites, n = 3 mice; p < 0.0001; Fig. 6c and e). When CSMN in UCHL1−/− mice were transduced by AAV2-UCHL1-IRES-eGFP, the integrity of apical dendrites was improved and the percent apical dendrites with vacuoles were dramatically reduced (32.41 ± 4.59%, n = 195 apical dendrites, n = 5 mice), becoming similar and comparable to that of WT mice (p = 0.2125). They were significantly lower than that of UCHL1−/− that are transfected by AAV2-eGFP alone (p < 0.0001).

Fig. 6:

Restoration of UCHL1 expression in only CSMN of UCHL1−/− mice is sufficient to improve their health and cytoarchitectural integrity.

a Schematic overview of the experimental setup. CSMN of wild type (WT), or UCHL1−/− mice were retrogradely transduced by injecting AAV2-eGFP into the corticospinal tract (CST) at P30, and CSMN apical dendrites were analyzed at P100. AAV2-UCHL1-IRES-eGFP virus was used to rescue UCHL1 expression in UCHL1−/− mice. b Representative images of GFP labeled CSMN after DAB mediated immunohistochemistry enhancement of the GFP signal allowing investigation of CSMN morphology in detail. Scale bar = 20 μm. c Primary apical dendrites of CSMN retrogradely labeled by AAV2-eGFP or AAV2-UCHL1-IRES-eGFP. Scale bar = 10 μm. d Cell bodies of CSMN retrogradely labeled by AAV2-eGFP or AAV2-UCHL1-IRES-eGFP. Scale bar = 20 μm. e Quantitative analysis of apical dendrites reveals significant decrease in the average percentage of apical dendrites with vacuoles when UCHL1 expression is restored to the UCHL1−/− CSMN. Data presented as mean ± SEM; ****p < 0.0001; one-way ANOVA with post hoc Tukey’s multiple comparison test. f Quantitative analysis of soma size shows significant increase in the average CSMN diameter when UCHL1 expression is restored to the UCHL1−/− CSMN. Data presented as mean ± SEM; **p < 0.01; one-way ANOVA with post hoc Tukey’s multiple comparison test.

In addition, while the soma diameter of CSMN in UCHL1−/− mice were significantly smaller (12.59 ± 0.22 μm, n = 457 cell bodies, n = 9 mice) than that of WT mice (15.47 ± 0.19 μm, n = 986 cell bodies, n = 5 mice; p = 0.0014; Fig. 6 d and f), when UCHL1 expression was selectively introduced to the CSMN of UCHL1−/− mice, soma diameter was also restored (14.85 ± 0.50 μm, n = 137 cell bodies, n = 5 mice), to the WT levels (p = 0.452).

Since AAV-mediated retrograde transduction targets only a small subgroup of CSMN, this application is more suitable for investigation of changes that occur in CSMN at a cellular level. As expected, we did not observe any improvement in overall motor function, which is assessed by rotarod, hanging wire and grip strength tests (Supplementary Fig. 2, Supplementary Table1).

UCHL1 expression restores cytoarchitectural integrity of CSMN with mSOD1 toxicity

Restoring UCHL1 expression in CSMN of UCHL1−/− mice is sufficient to improve their integrity. However, would UCHL1 expression also help improve the health of CSMN that degenerate due to other causes? CSMN of hSOD1G93A mice become vulnerable to degeneration very early in the disease and they display progressive degeneration with reduced soma size and apical dendrite degeneration [47-49]. To investigate whether overexpression of UCHL1 in CSMN that are diseased due to misfolded SOD1 would also improve their overall health, stability and cytoarchitectural integrity, we used the same AAV2 mediated gene delivery approach to deliver UCHL1 to CSMN that are diseased due to mutant SOD1 toxicity. The injections were performed at P60 and into the CST and CSMN were analyzed at P120 (Fig. 7a).

Fig. 7:

Expression of UCHL1 in CSMN of hSOD1G93A mice is sufficient to improve their health and cytoarchitectural integrity.

a Schematic overview of the experimental setup. CSMN of wild type (WT), or hSOD1G93A mice were retrogradely transduced by injecting AAV2-eGFP into the corticospinal tract (CST) at P60, and CSMN apical dendrites were analyzed at P120. AAV2-UCHL1-IRES-eGFP virus was used to express UCHL1 in CSMN of WT, or hSOD1G93A, mice. b Primary apical dendrites of CSMN retrogradely labeled by AAV2-eGFP or AAV2-UCHL1-IRES-eGFP. Scale bar = 20 μm. c Quantitative analysis of apical dendrites reveals significant decrease in the average percentage of apical dendrites with vacuoles when UCHL1 is expressed in CSMN of hSOD1G93A mice. Data presented as mean ± SEM; *p < 0.05, **p < 0.01, ****p < 0.0001; one-way ANOVA with post hoc Tukey’s multiple comparison test. d Cell bodies of CSMN retrogradely labeled by AAV2-eGFP or AAV2-UCHL1-IRES-eGFP. Scale bar = 20 μm. e Quantitative analysis of soma size shows significant increase in the average CSMN diameter when UCHL1 is expressed in CSMN of hSOD1G93A mice. Data presented as mean ± SEM; **p < 0.01, ****p < 0.0001; one-way ANOVA with post hoc Tukey’s multiple comparison test.

CSMN of WT mice displayed healthy robust apical dendrites (20.05 ± 4.33 %, n = 325 apical dendrites, n = 3 mice), whereas CSMN of hSOD1G93A mice had significantly higher percentages of vacuoles (58.90 ± 4.22 %, n = 341 apical dendrites, n = 4 mice; p < 0.0001; Fig. 7b–c). Overexpression of UCHL1 in CSMN of hSOD1G93A mice significantly improved cytoarchitectural integrity (16.18 ± 1.04 %, n = 442 apical dendrites, n = 6 mice; p < 0.0001) and was comparable to that of WT CSMN (p = 0.7394). Overexpression of UCHL1 was not toxic, as UCHL1 expression in WT CSMN further reduced percentage of vacuolated apical dendrites (4.31 ± 1.06 %, n = 189 apical dendrites, n = 4 mice; p = 0.0095).

Soma diameter of CSMN in hSOD1G93A mice were significantly smaller (12.25 ± 0.37 μm, n = 140 cell bodies, n = 4 mice) than that of WT mice (15.40 ± 0.41 μm, n = 94 cell bodies, n = 3 mice; p = 0.0012; Fig. 7d–e). When UCHL1 was overexpressed in the CSMN of hSOD1G93A mice, soma diameter was also restored (15.74 ± 0.40 μm, n = 155 soma, n = 7 mice, p < 0.0001), similar to the size of CSMN in WT mice (p = 0.9402). Expression of UCHL1 in healthy CSMN did not affect soma diameter (16.41 ± 0.41 μm, n = 85 soma, n = 5 mice, p = 0.4056).

UCHL1 expression in CSMN with TDP-43 pathology improves soma size and integrity of apical dendrite

TDP-43 pathology is one of the most common proteinopaties observed in a large population of ALS patients [84, 85]. Because cellular causes of CSMN degeneration are shared between UMNs in prpTDP-43A315T mice and ALS patients with TDP-43 pathology [51], and CSMN of prpTDP-43A315T mice display early CSMN loss [86], we next wanted to investigate whether expression of UCHL1 would also promote the health and integrity of CSMN that become diseased due to TDP-43 pathology, a pathology that is mostly excluded from ALS patients with SOD1 mutations [55, 56], at a cellular level. UCHL1 is selectively introduced to CSMN by AAV-mediated retrograde transduction at P60, an early symptomatic stage, and CSMN health is analyzed at P120 (Fig. 8a). We used the same cohort of WT mice as controls for both hSOD1G93A and prpTDP-43A315T mice. CSMN of prpTDP-43A315T mice contained significantly higher percentages of vacuolated dendrites compared to that of WT (68.34 ± 3.41 %, n = 243 apical dendrites, n = 4 mice; p < 0.0001; Fig. 8b–c). Expression of UCHL1 in CSMN of prpTDP-43A315T mice significantly improved cytoarchitectural integrity of apical dendrites (12.90 ± 1.63 %, n = 324 apical dendrites, n = 4 mice; p < 0.0001), such that they became comparable to that of WT CSMN (p = 0.3238). Soma diameter of CSMN in prpTDP-43A315T mice were significantly smaller (13.24 ± 0.39 μm, n = 99 soma, n = 4 mice, p = 0.0209; Fig. 8d–e). However, upon UCHL1 expression, the soma diameter of CSMN in prpTDP-43A315T mice were fully restored (17.29 ± 0.52 μm, n = 90 soma, n = 4 mice, p < 0.0001).

Fig. 8:

Expression of UCHL1 in CSMN of prpTDP-43A315T mice is sufficient to improve their health and cytoarchitectural integrity.

a Schematic overview of the experimental setup. CSMN of wild type (WT), or prpTDP-43A315T mice were retrogradely transduced by injecting AAV2-eGFP into the corticospinal tract (CST) at P60, and CSMN apical dendrites were analyzed at P120. AAV2-UCHL1-IRES-eGFP virus was used to express UCHL1 in CSMN of WT, or prpTDP-43A315T mice. b Primary apical dendrites of CSMN retrogradely labeled by AAV2-eGFP or AAV2-UCHL1-IRES-eGFP. Scale bar = 20 μm. c Quantitative analysis of apical dendrites reveals significant decrease in the average percentage of apical dendrites with vacuoles when UCHL1 is expressed in CSMN of prpTDP-43A315T mice. Data presented as mean ± SEM; *p < 0.05, ****p < 0.0001; one-way ANOVA with post hoc Tukey’s multiple comparison test. d Cell bodies of CSMN retrogradely labeled by AAV2-eGFP or AAV2-UCHL1-IRES-eGFP. Scale bar = 20 μm. e Quantitative analysis of soma size shows significant increase in the average CSMN diameter when UCHL1 is expressed in CSMN of prpTDP-43A315T mice. Data presented as mean ± SEM; *p < 0.05, ***p < 0.001, ****p < 0.0001; one-way ANOVA with post hoc Tukey’s multiple comparison test.

UCHL1 expression reduces both the levels of misfolded SOD1 and human TDP-43A315T

Since overexpression of the mutant form of SOD1 gene causes accumulation of misfolded SOD1 in hSOD1G93A mouse model of ALS, and this is one of the major underlying causes of CSMN degeneration, we next investigated whether UCHL1 expression would also reduce the levels of misfolded SOD1 in CSMN by using the well characterized B8H10 monoclonal antibody that can detect misfolded SOD1 protein in a wide spectrum of SOD1 mutants and metal-depleted WT SOD1 protein, but not intact WT SOD1 [49, 71]. CSMN of WT mice did not have misfolded SOD1, as expected (Fig. 9a). However, CSMN of hSOD1G93A mice contained high levels of misfolded SOD1 (1272 ± 26 arbitrary units (au), n = 844 CSMN, n = 6 mice vs WT: 224 ± 17 au, n = 1310 CSMN, n = 3 mice; p < 0.0001; Fig. 9a,b). AAV2-UCHL1-IRES-eGFP significantly reduced levels of misfolded SOD1in CSMN that are transduced with AAV expressing the UCHL1 gene (1123 ± 27 au, n = 255 CSMN, n = 5 mice; hSOD1G93A eGFP vs. hSOD1G93A UCHL1-IRES-eGFP p = 0.004; WT vs. hSOD1G93A UCHL1-IRES-eGFP p < 0.0001; Fig. 9a,b).

Figure 9:

a Representative images of CSMN and B8H10 antibody staining that recognizes misfolded SOD1 protein in the motor cortex of WT or hSOD1G93A mice with AAV2-eGFP or AAV2-UCHL1-IRES-eGFP injected into CST. Scale bar = 20 μm. b Average misfolded SOD1 fluorescence in CSMN of WT or hSOD1G93A mice with AAV2-eGFP or AAV2-UCHL1-IRES-eGFP injected into CST. c Representative images of CSMN and FLAG antibody staining that can be used to selectively visualize the N-terminus FLAG tag attached to the exogenous human TDP-43A315T protein in the motor cortex of WT or prpTDP-43A315T mice with AAV2-eGFP or AAV2-UCHL1-IRES-eGFP injected into CST. Scale bar = 20 μm. d Average human TDP-43A315T fluorescence in CSMN of WT prpTDP-43A315T mice with AAV2-eGFP or AAV2-UCHL1-IRES-eGFP injected into CST. Data presented as mean ± SEM; *p < 0.05, ***p < 0.001, ****p < 0.0001; one-way ANOVA with post hoc Tukey’s multiple comparison test.

PrpTDP-43 mice are generated by expressing the human TDP-43 protein with the A315T point mutation fused to an N-terminal FLAG tag [52], and it is mostly detected in the nucleus. To investigate the presence of human TDP-43A315T protein in CSMN of the prpTDP-43A315T mice, we used the FLAG monoclonal antibody. CSMN of WT mice do not have any human TDP-43A315T protein, as expected (Fig. 9c). However, CSMN of prpTDP-43A315T mice had high levels of human TDP-43A315T (1041 ± 20 arbitrary units (au), n = 1536 CSMN, n = 4 mice vs WT: 494 ± 5 au, n = 1239 CSMN, n = 3 mice; p < 0.0001; Fig. 9c,d). AAV2-UCHL1-IRES-eGFP significantly reduced levels of human TDP-43A315T, (866 ± 66 au, n = 111 CSMN, n = 3 mice; prpTDP-43A315T eGFP vs. prpTDP-43A315T UCHL1-IRES-eGFP p = 0.0257; WT vs. prpTDP-43A315T UCHL1-IRES-eGFP p = 0.0006; Fig. 9c,d).

DISCUSSION

Our results show proof for the cell-autonomous mechanisms responsible for CSMN degeneration and reveal that their demise is not correlated to or is a byproduct of SMN health. UCHL1 expression is required and sufficient to restore CSMN cytoarchitectural integrity and neuronal stability in UCHL1−/− mice, regardless of SMN health. Most interestingly, AAV-mediated gene delivery of UCHL1 directly to CSMN that become diseased due to mutant SOD1 toxicity or TDP-43 pathology, two distinct and non-overlapping causes of ALS in patients [55–58, 87, 88], also leads to significant improvement of CSMN stability, cytoarchitectural integrity and neuronal health, so much so that they become comparable to healthy control levels. Therefore, we not only show proof that improving UMN health in a cell-autonomous manner is a valid approach, but also identify UCHL1 as a potential candidate for gene therapy to diseased UMNs.

In an effort to bring a mechanistic insight for the CSMN degeneration in the absence of UCHL1 function, and to investigate whether introduction of UCHL1 would be sufficient to improve the cellular integrity of CSMN, we first knocked out UCHL1 protein selectively in the large subcerebral projection neurons (SCPN) in layer 5 or in SMN in the spinal cord by mating floxed UCHL1 (UCHL1f/f) mice with Rbp4cre or HB9cre mice respectively. Conditional knockout of UCHL1 only in SCPN, which had healthy SMN, replicated CSMN pathology observed in UCHL1−/− mice. Moreover, AAV2-mediated delivery of the UCHL1 only to CSMN was sufficient to improve their cytoarchitectural integrity in UCHL1−/− mice. Direct gene delivery of UCHL1 only to CSMN in UCHL1−/− mice, results in almost complete rescue of their neuronal stability, integrity of apical dendrites and spine, even when the rest of the mouse including SMN lack UCHL1 function. Therefore, we find that CSMN health or degeneration is not a function of SMN, and UMNs are indeed valid cellular targets for therapeutic interventions.

However, since the number of CSMN targeted by the AAV was kept intentionally small to allow individual analysis at a cellular level, it was not sufficient to show a functional improvement in motor behavior. Next step would be to deliver UCHL1 to as many CSMN as possible in the motor cortex either via direct cortex injection or by intrathecal injection, such that more effective and broader UCHL1 expression can be achieved. In these future studies the full extent of motor neuron circuitry with spinal cord and the neuromuscular junction needs to be investigated. Here, our goal was to focus on the cell biology of CSMN, and investigate the impact of UCHL1 expression on cellular stability and integrity. Now that we find UCHL1 treatment improves CSMN health and morphology, further detailed therapeutic studies are required.

UCHL1 is a unique protein. It has the ability to add and remove ubiquitin from proteins, playing an important role for their function, relocation within the cell, and ultimately determination of their fate for recycling [36–38, 89]. UCHL1 function is required to ensure free ubiquitin levels are stable in cells, especially in neurons. The ubiquitin–proteasome system plays a fundamental role in maintaining protein homeostasis, which is a crucial factor in the development of motor neuron diseases [90]. Even though all neurons ubiquitously express UCHL1, the UMNs have high levels of UCHL1 throughout life.

Interestingly, mutations in the UCHL1 gene, which is located on chromosome 4p13, resulted in numerous forms of neurodegeneration that affects movement. For example, autosomal recessive spastic paraplegia-79 (SPG79) is caused by homozygous or compound heterozygous mutation in the UCHL1 gene [42]. UCHL1GLU7ALA missense mutation lies within the ubiquitin-binding domain of UCHL1 protein and leads to loss of hydrolase function [39]. Siblings homozygous for the mutation have spasticity with UMN dysfunction. Patients with other missense mutations, such as UCHL1ARG178GLN and UCHL1ALA216ASP developed spasticity and ataxia following child onset blindness [41]. The UCHL1ALA216ASP mutation was reported to be insoluble and nonfunctional, whereas the UCHL1ARG178GLN mutation leads to a 4-fold increase in the hydrolytic activity. Recently, a third family with two siblings carrying a deleterious homozygous splice-site variant had spasticity [40]. Two siblings with Gly210del deletion had visual impairment, progressive spasticity, weakness, atrophy of the lower legs and ataxia [43]. Clinical features of all these 10 patients with mutations in their UCHL1 gene share early neurodegeneration with indications of UMN involvement.

Similar to patients with mutations in their UCHL1 gene, mouse models of UCHL1, especially the Uchl1 (UCHL1−/−) mice used in this study lack all UCHL1 function and display early onset neurodegeneration, accompanied with spasticity, muscular atrophy and profound upper motor neuron degeneration with disintegrating apical dendrites and spine loss [44-46]. Therefore, we think UCHL1−/− mice offers a unique opportunity to investigate upper motor neuron degeneration.

Numerous mouse models are generated to mimic ALS disease pathology. The hSOD1G93A and the prpTDP-43A315T models are one of the best characterized mouse models of ALS, recapitulating many aspects of disease pathology [47, 48, 50–52, 80–83]. TDP-43 pathology is observed in the brains of about 95% of ALS patients [84] regardless of a mutation in the TARDP gene [85, 91] and is distinct from fALS patients with SOD1 mutations [87]. Patients with SOD1 mutations or the SOD1 mouse models do not have TDP-43 pathology in their brains [55–58, 88]. Therefore, these two models represent two distinct disease mechanisms and a broad spectrum of patients. Interestingly, CSMN of these mice undergo progressive degeneration, which starts early with apical dendrite degeneration and spine loss [70]. Regardless of specific genes and mutations, disruption of protein homeostasis and canonical pathways such as ER stress, protein ubiquitination and unfolded protein response seem to be a common underlying cause for ALS and other neurodegenerative diseases [92-95]. It is possible UCHL1 provides benefit in both hSOD1G93A and the prpTDP-43A315T models due to its important role in the UPS pathway clearing toxic misfolded proteins [36, 37, 96, 97].

CSMN are an important member for the cortical component of motor neuron circuitry [4–6, 12]. While their soma resides in layer 5 of the motor cortex, their apical dendrite extend towards upper layers and the axon innervate spinal cord targets, forming a connection between brain and the spinal [2]. Their apical dendrite is a critical site for proper modulation. Therefore, the integrity and the stability of the apical dendrite is paramount for UMN health and function [2, 48, 98, 99].

We recently discovered that Betz cells of sporadic ALS, familial ALS and ALS-Frontotemporal Dementia patients display massive apical dendrite degeneration [1, 9, 100], and this cellular pathology is shared among species and observed in CSMN of numerous well characterized mouse models of ALS, such as hSOD1G93A [47, 48, 70], TDP-43A315T [51, 52, 80, 83], hPFN1G118V [54], and AlsinKO mice [53]. Therefore, identification of a treatment strategy that improves cytoarchitectural integrity, apical dendrite stability and eliminate spine loss would be translational. In addition, investigation of CSMN’s requirements for survival and improved health is of great importance, as this information is clinically relevant, and has the potential to exponentially improve the success rates of future clinical trials of motor neuron diseases in which upper motor neurons and the motor neuron circuitry are affected.

UCHL1 protein fused to the protein transduction domain of HIV-transactivating transduction protein (TAT-UCHL1) can transduce neurons after intraperitoneal (i.p.) injection into mice [101]. In a controlled cortical impact (CCI) injury model of post-traumatic brain injury, TAT-UCHL1 treatment improved function of the ubiquitin-proteasome pathway, decreased activation of autophagy after CCI, attenuated axonal injury and increased hippocampal neuronal survival after CCI [101]. During hypoxic injury, whereas pharmacologic inhibition of UCHL1 function exacerbates neuronal death induced by hypoxia, TAT-UCHL1 treatment provides neuroprotection [102]. In a mouse model of AD, TAT-UCHL1 restored synaptic function in hippocampal slices after oligomeric Aβ treatment, and improved retention of contextual learning in APP/PS1 mice upon i.p. injection [103]. Moreover, Aβ induced impairment of neurotrophin-mediated retrograde signaling can be rescued by increasing cellular UCHL1 levels upon TAT-UCHL1 treatment [104]. These, in combination with our UCHL1 rescue data using AAV-mediated gene delivery of UCHL1 directly to CSMN suggest that UCHL1 could potentially be a therapeutic agent for treatment of not only a wide variety of neurodegenerative diseases, but also other injuries such as TBI, hypoxia, and cerebral ischemia.

Taken together, our results reveal that CSMN degeneration occurs independently of SMN health. UMNs, which play a pivotal role for the initiation and modulation of movement, are indeed cellular targets for therapeutic interventions. We also report UCHL1 as a potential candidate for gene therapy approaches, especially for diseased UMNs. Because UMNs in mice and UMNs in patients share the same underlying pathologies, building evidence suggests that improving the integrity and stability of UMNs at a cellular level will offer a therapeutic benefit for the motor neuron circuitry that degenerates in patients. Since mutant SOD1 toxicity and especially TDP-43 pathology are observed in a wide spectrum of ALS patients, our results suggest that directed gene delivery of UCHL1 to diseased UMNs could offer a novel therapeutic intervention strategy for a broad spectrum of diseases, such as ALS, HSP, and PLS.